Plasmid 65 (2011) 118–124

Contents lists available at ScienceDirect

Plasmid

journal homepage: www.elsevier .com/ locate /yplas

Streptomyces turgidiscabies Car8 contains a modular pathogenicityisland that shares virulence genes with other actinobacterial plantpathogens

José C. Huguet-Tapia a, Jonathan H. Badger b, Rosemary Loria a,⇑, Gregg S. Pettis c,⇑a Department of Plant Pathology and Plant-Microbe Biology, Cornell University, Ithaca, NY, USAb The J. Craig Venter Institute, San Diego, CA, USAc Department of Biological Sciences, Louisiana State University, Baton Rouge, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 23 August 2010Accepted 8 November 2010Available online 16 November 2010Communicated by Dr. J. Casadesus

Keywords:VirulenceGenomic islandHorizontal gene transferPotato scabStreptomyces

0147-619X/$ - see front matter � 2010 Elsevier Incdoi:10.1016/j.plasmid.2010.11.002

⇑ Corresponding authors. Addresses: 334 PlanDepartment of Plant Pathology and Plant-MicroUniversity, Ithaca, NY 14850, USA. Fax: +1 607 255Life Sciences Building, Department of Biological ScieUniversity, Baton Rouge, LA 70803, USA. Fax: +1 225

E-mail addresses: RL21@cornell.edu (R. Loria),Pettis).

Streptomyces turgidiscabies Car8 is an actinobacterium that causes the economically impor-tant disease potato scab. Pathogenesis in this species is associated with a mobile pathoge-nicity island (PAISt) that site specifically inserts into the bacA gene in Streptomyces spp.Here we provide the 674,223 bp sequence of PAISt, which consists of two non-overlappingmodules of 105,364 and 568,859 bp. These modules are delimited by three copies of an8 bp palindromic sequence (TTCATGAA), that also is the integration site (att) of the ele-ment. Putative tyrosine recombinase (IntSt) and excisionase (XisSt) proteins are encodedjust upstream of att-R. PAISt has regions of synteny to pathogenic, symbiotic and sapro-phytic actinomycetes. The 105,364 bp PAISt module is identical to a genomic island inStreptomyces scabies 87–22, while the 568,859 bp module contains only a short region ofsynteny to that genome. However, both modules contain previously characterized and can-didate virulence genes.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction

Acquisition of accessory genes via horizontal genetransfer (HGT) is one of the driving forces in bacterial evo-lution. Genes are often mobilized in syntenic blockstermed genomic islands (GEIs). GEIs are distinct regionsof DNA that have, or once had, functional mobility genes.This category of elements includes many types of mobileelements: integrative and conjugative elements (ICEs),conjugative transposons, and some prophages (reviewedin Juhas et al. (2009)). The term GEI refers to relatively

. All rights reserved.

t Science Building,be Biology, Cornell4471 (R. Loria); 202nces, Louisiana State578 2597 (G. Pettis).

gpettis@lsu.edu (G.S.

large regions (>20 kb) that differ among microbial strainsand share features derived from their evolutionary history,including variation in nucleotide bias, insertion in tRNAgenes, flanking direct repeats, and functional or crypticmobility genes. Importantly, GEIs contribute to genomeplasticity and strain differentiation by providing new func-tions such as antibiotic production, antibiotic resistance,catabolic activities, or virulence to the recipient strain.The role of HGT in the evolution of plant and animal viru-lence in microbes has been known for some time and GEIsthat confer virulence are known as pathogenicity islands(PAIs). Though the concept of PAIs was introduced morethan 10 years ago (Groisman and Ochman, 1996), exam-ples of PAIs in plant pathogens are almost entirely limitedto Gram-negative bacteria.

The first example of a PAI in a Gram-positive plant path-ogen was from the root and tuber pathogen Streptomycesturgidiscabies (Kers et al., 2005), PAISt. This emergent path-ogen was first discovered on the island of Hokkaido, Japan,

J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124 119

where it causes disease in commercial potato fields (Miyajimaet al., 1998; Takeuchi et al., 1996). DNA sequence of a smallpart of PAISt revealed that it carries genes encoding at leastfour virulence factors: the biosynthetic pathway for thephytotoxin thaxtomin (Txt proteins), a secreted saponinase(TomA), a secreted necrogenic protein (Nec1), and a cytoki-nin biosynthetic pathway (Fas proteins) (Kers et al., 2005).Surprisingly, this study also demonstrated that PAIStmobilizes from S. turgidiscabies and integrates into thechromosome of nonpathogenic Streptomyces spp. includingS. coelicolor, S. diastochromogenes and S. lividans. In the caseof S. diastatochromogenes, acquisition of PAIst was suffi-cient to confer a pathogenic phenotype on the organism.Integration of PAISt occurs at the palindromic sequenceTTCATGAA located within the 30 end of the bacitracin resis-tance gene (bacA) in S. turgidiscabies Car8 and all recipientstrains (Kers et al., 2005).

Streptomycete PAISt transconjugants typically containthe approximately 660 kb element, or a truncated versionof approximately 100 kb (Kers et al., 2005). These data sug-gest that PAISt is a very large GEI that has the ability tomobilize and site specifically insert during conjugation,meeting the criterion of an ICE. However, two propertiesof PAISt mobilization are unusual relative to describedICEs: these elements typically insert into the 30 end of tRNAencoding genes and do not mobilize in a modular fashion(Burrus et al., 2002). Here we provide the complete se-quence of PAISt and an analysis of gene content relativeto the pathogenic phenotype and mobility.

2. Materials and methods

2.1. Annotation of the PAISt

A pseudomolecule representing the chromosome of S.turgidiscabies Car8 was built from the scaffolds of the S. tur-gidiscabies Car8 genome-sequencing project (GenBankAccession No.: AEJB00000000). MUMmer 3.0 software(Kurtz et al., 2004) was used to order the scaffolds usingthe finished sequence of the S. scabies 87–22 chromosomeas a reference (GenBank Accession No.: FN554889). Previ-ously published sequence data from the PAISt (Kers et al.,2005) was used to delimit the boundaries of the PAISt inthe pseudomolecule. The landmark sequences included inthe analysis were: nec1, the txt genes, and the 30 end ofbacA (Genbank Accession Nos.: AY707080, AY707081 andAY707082, respectively). The 674,223 bp PAI sequencewas initially fragmented into 18 contigs with numerouslow coverage regions. The relevant sequence was closedto a single, high quality finished contig using 149 Sangerclone primer walks and resulted in a 717,473 bp contig,which included the closed PAI sequence.

Coding sequences in the PAISt were predicted usingGLIMMER 3.2 (Delcher et al., 1999) trained with S. scabies87–22 genes. Predicted genes were automatically anno-tated using Blast2GO (Conesa and Gotz, 2008). Manualannotation was performed using the Artemis genomeviewer (Rutherford et al., 2000), and the Gene Ontology(2006) and Pfam (Bateman et al., 2004) databases.

2.2. Comparative sequence analysis

Predicted coding sequences were compared at the ami-no acid level with the genomes of selected actinobacteriaand plant pathogenic bacteria (Supplementary Table S1).Homologs were identified with BLAST (Altschul et al.,1990), using the e-value of 0.0001 as a threshold for signif-icance, and identity scores were used to build a matrix ofbest hits. This matrix was used to generate a color-codedBLAST-fingerprinting plot with the R statistical package(http://www.r-project.org/). Additionally, sequence com-parisons and analysis of gene synteny were carried outusing the Artemis Comparison Tool (Carver et al., 2005).

3. Results and discussion

3.1. Modular nature of the PAI

The complete DNA sequence of PAISt was extractedfrom a draft genome sequence of S. turgidiscabies Car8(GenBank Accession No.: AEJB00000000). The sequencedata revealed that PAISt is an integrated genetic elementof 674,223 bp containing two non-overlapping modulesof 105,364 and 568,859 bp (Fig. 1, Supplementary DataS2). These data agree with the experimentally estimatedsizes (660 and 100 kb) of the versions of PAISt transferredfrom S. turgidiscabies Car8 to S. coelicolor (Kers et al., 2005).Three copies of the 8 bp palindrome sequence TTCATGAAdelimited the two modules of PAISt (Fig. 1); this sequenceis also the experimentally demonstrated genomic integra-tion site (att-site) of the element (Kers et al., 2005). att-Lwas located within the 30 region of bacA (1433,070 bp onthe S. turgidiscabies Car8 pseudomolecule), while att-I(1538,434 bp) and att-R (2107,293 bp) were each embed-ded in a copy of the 30 end of bacA, suggesting a conservedrecombination event. The average G + C content of PAIStwas 68%, which is somewhat lower than that typical ofstreptomycete genomes (72%). PAISt is predicted to con-tain 647 coding sequences (CDS) (Supplementary DataS2) with an average length of 883 bp.

3.2. Tyrosine integrase and excisionase candidates

Previous studies demonstrated that PAISt transfers fromS. turgidiscabies Car8 by conjugation to other Streptomycesspp. and integrates through a site-specific recombinationmechanism (Kers et al., 2005), consistent with the hypoth-esis that PAISt is an ICE. Typically, an ICE encodes a tyro-sine-recombinase adjacent to an att site (Burrus andWaldor, 2004; Ramsay et al., 2006; Scott and Churchward,1995). The CDS stPAI0647 lies just upstream of att-R (Fig. 1,Supplementary Data S2). PSI-BLAST (three iterations) ofthe 466 amino acid sequence against the non-redundantdatabase in Genbank indicated that stPAI0647 is 18% iden-tical to a tyrosine recombinase found in the firmicuteAllicyclobacillus acidocaldarius (GenBank Accession No.:ZP_03494792). There are five additional putative integras-es on PAISt (stPAI0447c, stPAI0567, stPAI0570, stPAI0575,and stPAI0632) but none are located near att sites(Supplementary Data S2). Interestingly, a copy of the att

10.8 MbS. turgidiscabies Car8 pseudomolecule

105,364 bp

1 433,070 bp

568,859 bp

stPAI0643 stPAI0644 stPAI0645 stPAI0646 stPAI0647

Bacitracin resistance gene (bacA) or its truncated 3’ end

Regulator

NAD dependent epimerase

Hydrolase

Excisionase

Integrase att-R

att-R att-I att-L

Fig. 1. The characteristics of PAISt and its location in the pseudomolecule of S. turgidiscabies Car8. PAISt is organized as a complex ICE within thechromosome of S. turgidiscabies Car8 (top portion of the figure). The 8 bp integration (att) sites are imbedded within the duplication of the 30 end of bacA(black arrows in the middle portion). A putative integrase (stPAI0674; IntSt) and a putative excisionase (stPAI0646; XisSt) are encoded adjacent to att-R(bottom portion). Upstream from IntSt and XisSt is a gene encoding a putative transcriptional regulator (stPAI0643) that could be involved in regulatingexcision and integration.

120 J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124

site is imbedded in the C-terminal region of stPAI0570;however, there is no evidence of recombination (duplicationof the 3’ end of the bacitracin resistance gene). These datasuggest, but do not prove, that stPAI0647 (IntSt) serves tosite specifically integrate PAISt into the att site in the bacAgene. It is worth noting that tyrosine recombinases areresponsible for integration and excision of a wide array ofintegrative elements and usually, but not exclusively, cata-lyze integration into the 30 end of tRNA genes (9).

ICE integrases often function in conjunction with exci-sionase proteins (Xis) which share little homology at theamino acid level but are typically small proteins with basicisoelectric points (Lee et al., 2007). Adjacent to IntSt thereis an ORF (stPAI0646) that encodes a putative 118 aminoacid protein with a basic isoelectic point (13.8 kDa; pI11.76). Based on the location, size and pI of the protein en-coded by stPAI0646, it is reasonable to suggest that thisgene encodes a protein (XisSt) that functions in the exci-sion of PAISt (Supplementary Data S2).

Bioinformatic data presented here, in combination withexperimental data presented previously (Kers et al., 2005),suggest that IntSt and XisSt are responsible for the excisionand integration of the entire PAISt (674,223 bp) and thesmaller module (105,364 bp). However, it may be thatthe whole element is transferred and only the smallermodule is stably integrated, in some cases. Based on theorganization of PAISt, it is expected that the 568,859 bpmodule (Fig. 1) also should be integrated into recipientgenomes. However, in experiments demonstrating integra-tion, the selectable marker was located within the105,364 bp module (at nec1), therefore transfer of the568,859 bp module would not have been detected (Kerset al., 2005). Regardless, the possibility that IntSt and XisStmight be able to excise and integrate multiple modules ofPAISt is noteworthy and worthy of experimental follow up.

3.3. Characteristics of the 105,364 bp PAI module

Interestingly, the entire 105,364 bp PAISt module isidentical to a nonmobile GEI (PAISs1) located in the chro-mosome of the plant pathogen S. scabies 87–22 (Figs. 2and 3). As in S. turgidiscabies Car8, PAISs1 is integrated intothe 30 end of the bacA gene (SCAB77601) in S. scabies 87–22. Importantly, the att site at the 30 end of the S. scabies87–22 PAISs1 is mutated from TTCATGAA to TGTATGAAand contains a degenerated version of IntSt (SCAB76721;Genbank accession YP_003493179). It is possible that themutation in the att site of PAISs1 and the gene erosion ofthe integrase cause the fixation of this GEI in S. scabies87–22.

The CDSs stPAI001 – stPAI0081 constitute the105,364 bp module encoded in S. turgidiscabies Car8(Supplementary Data S2). Most of the highest scoringBLAST hits are from the genus Streptomyces, but a numberof other actinomycetes are represented. The large numberof putative secreted hydrolytic proteins, particularly gly-coside hydrolases, in this module is noteworthy and con-sistent with the degradation of plant tissue caused byplant pathogenic streptomycetes (Loria et al., 2006). Pre-dicted hydrolase genes were interspersed with putativeABC transporter genes, some of which are predicted toencode sugar transporters. Many regulators, includingLacI, IcIR, TetR, MarR, XRE, PadR, and two-componentfamilies, are also encoded in this region. There are a totalof five LacI and IcIR family regulators in the 105,364 mod-ule (Supplementary Data S2), consistent with their role inregulating carbohydrate metabolism (Nguyen and Saier,1995).

Encoded in this region is TomA, a putative secreted vir-ulence protein that has been characterized as a tomatinasein S. scabies (Seipke and Loria, 2008) and is known to be

10.1 Mb

104 kb

18.3 kb

S. scabies 87.22:

TTCATGAA TGTATGAA

6, 531,990 bp

PAISs2

PAISs1

nos P450 txtR txtA txtB txtC

nos P450 txtR txtA txtB txtC

90 87 78 90 93 89

PAISt

PAISs2

Nitric oxide synthase (nos)

P450 monooxygenase (P450 and txtC)

Regulator (txtR)

Thaxtomin synthase (txtA and txtB)

Insertion sequence (IS)

Unknown

SISI IS

Fig. 2. The GEIs in S. scabies 87–22 and syntenic regions in the S. turgidiscabies PAIst and S. scabies PAISs2. The S. scabies 87–22 chromosome (top portion ofthe figure) contains the GEIs PAISs1 (104 kb) and PAISs2 (18.3 kb). The 30 end of PAISs1 has a degenerate integration site (underlined). The thaxtominbiosynthesis genes are encoded in syntenic regions in the S. turgidiscabies Car8 PAISt and S. scabies 87–22 PAISs2 (bottom portion). White bars indicatehomology and percentage identity at the amino acid level. The predicted functions of genes in the depicted regions are provided in the legend.

J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124 121

conserved in the pathogen S. acidiscabies (Loria, 2006). Thisprotein also has a homolog in the actinobacterial plantpathogen Clavibacter michiganensis subsp. michiganensis(CMM) (Gartemann et al., 2008), but is better known as avirulence factor in plant pathogenic fungi (Ito et al.,2004; Martin-Hernandez et al., 2000). Interestingly, TomAis embedded in a region (stPAI0021–stPAI0035) in whichnine out of 15 CDSs encode proteins with best hits toCMM (Supplementary Data S2). The CDSs stPAI0028, andstPAI0031–stPAI0035 are syntenous in the two genomesand are predicted to encode two beta-glycosidases, anABC sugar transporter, and a TetR family regulator. Thesedata suggest the hypothesis that CMM and plant patho-genic streptomycetes share a virulence mechanism involv-ing carbohydrate metabolism.

The secreted protein Nec1 is another characterized vir-ulence protein that is encoded on the 105,364 bp module(Supplementary Data S2) (Bukhalid and Loria, 1997; Kerset al., 2005). Nec1 lacks homologs in the public databasesoutside plant pathogenic streptomycetes and has no char-acterized motifs, yet has a strong virulence phenotype oninoculated plant roots (Joshi et al., 2007a). The DNA se-quence of nec1 is almost identical in diverse plant patho-genic streptomycetes, but has a particularly low G + Ccontent, consistent with very recent horizontal transferamong pathogens (Bukhalid et al., 2002). A previous studyused Southern analysis to demonstrate conservation of thenec1 gene and a 26 kb flanking region across plant patho-genic streptomycetes in the diastatochromogenes cluster(Bukhalid et al., 2002); these data are consistent with thesequence analysis provided in this study.

3.4. Characteristics of the 568,859 PAI module

The CDSs stPAI0082–stPAI0647 constitute the568,859 bp module in S. turgidiscabies Car8 (Fig. 1, Supple-mentary Data S2). Overall, this module is characterized bya preponderance of hypothetical proteins, which is notunusual for ICEs. However, a region of synteny to theS. scabies 87–22 genome also exists in this module (Fig. 2,Fig. 3). This syntenic region is a GEI in S. scabies 87–22(PAISs2) that encodes six characterized thaxtomin A bio-synthetic and regulatory genes (Healy et al., 2002; Healyet al., 2000; Johnson et al., 2009; Joshi et al., 2007b). A re-gion of synteny to the Rhodococcus fascians D188 genomelies adjacent to the thaxtomin biosynthetic pathway inS. turgidiscabies and is composed of the previously de-scribed fas operon (Supplementary Data S2, fas1–5); thesevirulence genes are not present in S. scabies 87–22 (Joshiand Loria, 2007).

The thaxtomin and fas biosynthetic pathways are theonly characterized virulence genes on the larger PAI mod-ule. However, the 568,859 bp PAISt module contains anumber of gene clusters with synteny to other actinomy-cetes (Fig. 3). An interesting cluster of five genes(stPAI0309–stPAI0313) encodes proteins syntenic to theplant symbiotic actinobacterium Frankia sp. CcI13 thatmay constitute a metabolic pathway (Supplementary DataS2). Another interesting region in the PAISt is a putativelantibiotic biosynthetic pathway (stPAI0324, stPAI0325,stPAI0326 and stPAI0327). This region is syntenic with apredicted biosynthetic operon in the saprophyticstreptomycete S. griseus (Supplementary Data S2). Since

Fig. 3. BlastP-fingerprinting plot of the PAISt. Coding sequences of PAISt are denoted by the horizontal line at the bottom of the plot (0–675 kb). Thepredicted proteome of PAISt was compared to predicted proteomes of selected bacteria (primarily actinobacteria) listed on the right side of the plot. SeeTable S1 for complete names and genome accession numbers. Percentage identity is indicated by colors shown in the legend at the bottom of the figure.

122 J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124

lantibiotics are active against other Gram-positive bacteria(McAuliffe et al., 2001; Willey and van der Donk, 2007),this pathway might provide an advantage to S. turgidisca-bies Car8 in rhizosphere colonization or in competing fornutrients in necrotized plant tissue. Discrete regions atthe 30 end of PAISt displayed high identity with the sapro-phytes S. avermitilis and S. coelicolor (Fig. 3). Alignment ofPAISt with the S. avermitilis and S. coelicolor chromosomesrevealed syntenic clusters of genes coding for regulatorsand transporter systems (Fig. 4, Supplementary Data S2).Interestingly, these regions are located on the arms of boththe S. avermitilis and S. coelicolor chromosomes, suggestingthat their functions are niche specific (Bentley et al., 2002).

3.5. Comparison with actinomycete ICEs

Previously described actinomycete ICEs share homologsthat function in excision and integration, replication, con-jugative transfer, and regulation, as well as a structuralorganization (te Poele et al., 2008a,b). Actinomycete ICEshave the ability to excise from the chromosome in donorcells, replicate autonomously like a plasmid, transfer to

recipient cells as double-stranded DNA and, finally, inte-grate into the recipient’s chromosome (te Poele et al.,2008a,b). In PAISt, IntSt, XisSt and a putative transcrip-tional regulatory gene (stPAI0643) cluster at an att site(Fig. 1), which is consistent with the organization of otheractinomycete ICEs (te Poele et al., 2008a,b). However,conjugation, replication and regulatory proteins typicallyassociated with actinomycete ICEs and encoded near theint and xis genes, are lacking in PAISt (Supplementary DataS2). There are 11 genes clustered within the regionstPAI106–stPAI185 that are syntenic to genes on an inte-grated plasmid in the actinobacterium Corynebacteriumglutamicum ATCC 13032 (Supplementary Fig. S3), whichis a species used to produce amino acids in large-scale fer-mentation (Hermann, 2003). This region of synteny dis-plays a peculiar distribution of several genes that encodeputative DNA-processing proteins that could participatein mobilization or replication of PAISt (SupplementaryData S2). However, none of these proteins are homologouswith the main transfer protein (TraB) of actinomyceteICE elements and Streptomyces plasmids. TraB proteinsare similar to the FtsK-SpoIIIE family of septal DNA

Fig. 4. Synteny of PAISt with multiple two-component response regulators and ABC transporters genes in S. coelicolor and S. avermitilis. White bars indicatehomology and percentage identity at the amino acid level. The predicted functions of genes are provided in the legend.

J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124 123

translocases and are responsible for transferring actinobac-terial plasmids intercellularly in the novel double-strandedform mentioned earlier (te Poele et al., 2008a). With thecaveat that conjugation functions may yet be identifiedon PAISt, the results here raise the possibility that PAISttransmission may instead occur via mobilization by anadditional undetermined conjugal element present inS. turgidiscabies. If PAISt is in fact found to be self-transmis-sible, the lack of TraB homologs would suggest that themechanism of PAISt conjugal transfer will differ signifi-cantly from that of other actinomycete ICEs.

3.6. Possible recombination events

DNA sequence comparisons suggest at least two possi-ble scenarios for recombination events that led to the for-mation of the PAIs in S. turgidiscabies and S. scabies (Fig. 1).It may be that the PAISt was originally an ICE in S. scabiesthat progressively acquired novel DNA. The entire islandcould have been transferred directly from S. scabies, orvia other donors, to S. turgidiscabies. In S. scabies, gene ero-sion and recombination might have resulted in the loss ofthe recombinase, conjugation genes, and the fas operon, aswell as the degeneration of the integration sites. Subse-quently, the PAI might have undergone recombination

events that separated the island into two modules, PAISs1(containing nec1 and tomA) and the PAISs2 (containing thethaxtomin biosynthetic cluster), fixing this element in thebacterium. Furthermore, the functional integrase on PAIStmay have served in the acquisition of DNA into the ele-ment (Biskri et al., 2005; Fonseca et al., 2008).

Another possible scenario is that S. turgidiscabiesacquired a small version of the island from S. scabies(105,364 bp PAI module). The thaxtomin biosynthetic clus-ter and fas operon could have been acquired subsequentlyduring independent recombination events, forming geno-mic ‘‘islets’’ within the PAISt. Transposons and ISs in thevicinity of the thaxtomin genes and fas operon suggest thatthese regions might have been acquired through transposi-tion. Of course, neither of these scenarios addresses the in-tra species genomic variability in these species, a subject offuture investigations.

4. Conclusions

To the best of our knowledge, PAISt is the largest mobileGEI described to date. It is very likely that PAISt has playeda pivotal role in the evolution of S. turgidiscabies. In fact, itis tempting to speculate that acquisition of PAISt providedthe genetic potential for S. turgidiscabies to develop a

124 J.C. Huguet-Tapia et al. / Plasmid 65 (2011) 118–124

pathogenic relationship with plants. This PAI containsmost of the characterized virulence genes in plant patho-genic streptomycetes (Loria et al., 2006). Thaxtomin A, apotent cellulose biosynthesis inhibitor, is required for viru-lence in scab-causing streptomycetes and likely functionsto facilitate inter and intracellular penetration of the planthost by this filamentous pathogen. Interestingly, thaxto-mins are only produced by plant pathogenic streptomy-cetes. The fas genes on PAISt are homologous to the fasoperon in the actinobacterial plant pathogen R. fasciansand encode biosynthesis of cytokinins that have a viru-lence phenotype in both pathogens (Joshi and Loria,2007). The secreted protein Nec1 has a dramatic virulencephenotype in root infection and is highly conserved amongplant pathogenic streptomycetes. The secreted proteinTomA, also highly conserved among plant pathogenicstreptomycetes, has homologs in the actinobacterial plantpathogen CMM (Seipke and Loria, 2008). The PAISt lacksthe biosynthetic pathway for a coronafacic acid-like bio-synthetic cluster that contributes to virulence in S. scabies87–22 (Bignell et al., 2010). Furthermore, there are manyexcellent virulence gene candidates on PAISt.

Acknowledgments

This work was supported by the National ResearchInitiative of the USDA Cooperative State Research, Educationand Extension Service, Grant Nos. 2007-35600-17813 (toG. Pettis) and 2010-65110-20416 (to R. Loria).

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.plasmid.2010.11.002.

References

Altschul, S.F. et al., 1990. Basic local alignment search tool. J. Mol. Biol.215, 403–410.

Anonomous, 2006. The Gene Ontology (GO) project in 2006. Nucleic AcidsRes. 34, D322–D326.

Bateman, A. et al., 2004. The Pfam protein families database. Nucleic AcidsRes. 32, D138–D141.

Bentley, S.D. et al., 2002. Complete genome sequence of the modelactinomycete Streptomyces coelicolor A3(2). Nature 417, 141–147.

Bignell, D.R.D. et al., 2010. Streptomyces scabies 87–22 contains acoronafacic acid-like biosynthetic cluster that contributes to plant-microbe interactions. Mol. Plant–Microbe Interact. 23, 161–175.

Biskri, L. et al., 2005. Comparative study of class 1 integron and Vibriocholerae superintegron integrase activities. J. Bacteriol. 187, 1740–1750.

Bukhalid, R.A., Loria, R., 1997. Cloning and expression of a gene fromStreptomyces scabies encoding a putative pathogenicity factor. J.Bacteriol. 179, 7776–7783.

Bukhalid, R.A. et al., 2002. Horizontal transfer of the plant virulence gene,nec1, and flanking sequences among genetically distinct Streptomycesstrains in the diastatochromogenes cluster. Appl. Environ. Microbiol.68, 738–744.

Burrus, V. et al., 2002. Conjugative transposons: the tip of the iceberg.Mol. Microbiol. 46, 601–610.

Burrus, V., Waldor, M.K., 2004. Shaping bacterial genomes withintegrative and conjugative elements. Res. Microbiol. 155, 376–386.

Carver, T.J. et al., 2005. ACT: the Artemis comparison tool. Bioinformatics21, 3422–3423.

Conesa, A., Gotz, S., 2008. Blast2GO: A Comprehensive suite for functionalanalysis in plant genomics. Int. J. Plant Genom. 2008, 619832.

Delcher, A.L. et al., 1999. Improved microbial gene identification withGLIMMER. Nucleic Acids Res. 27, 4636–4641.

Fonseca, E.L. et al., 2008. New qnr gene cassettes associated withsuperintegron repeats in Vibrio cholerae O1. Emerg. Infect. Dis. 14,1129–1131.

Gartemann, K.H. et al., 2008. The genome sequence of the tomato-pathogenic actinomycete Clavibacter michiganensis subsp.michiganensis NCPPB382 reveals a large island involved inpathogenicity. J. Bacteriol 190, 2138–2149.

Groisman, E.A., Ochman, H., 1996. Pathogenicity islands: bacterialevolution in quantum leaps. Cell 87, 791–794.

Healy, F.G. et al., 2002. Involvement of a cytochrome P450monooxygenase in thaxtomin A biosynthesis by Streptomycesacidiscabies. J. Bacteriol. 184, 2019–2029.

Healy, F.G. et al., 2000. The txtAB genes of the plant pathogen Streptomycesacidiscabies encode a peptide synthetase required for phytotoxinthaxtomin A production and pathogenicity. Mol. Microbiol. 38, 794–804.

Hermann, T., 2003. Industrial production of amino acids by coryneformbacteria. J. Biotechnol. 104, 155–172.

Ito, S. et al., 2004. Tomatidine and lycotetraose, hydrolysis products ofalpha-tomatine by Fusarium oxysporum tomatinase, suppress induceddefense responses in tomato cells. FEBS Lett. 571, 31–34.

Johnson, E.G. et al., 2009. 4-nitrotryptophan is a substrate for the non-ribosomal peptide synthase TxtB in the thaxtomin A biosyntheticpathway. Mol. Microbiol. 73, 409–418.

Joshi, M.V. et al., 2007a. Streptomyces turgidiscabies secretes a novelvirulence protein, Nec1, which facilitates infection. Mol. Plant–Microbe Interact. 20, 599–608.

Joshi, M.V. et al., 2007b. The AraC/XylS regulator TxtR modulatesthaxtomin biosynthesis and virulence in Streptomyces scabies. Mol.Microbiol. 66, 633–642.

Joshi, M.V., Loria, R., 2007. Streptomyces turgidiscabies possesses afunctional cytokinin biosynthetic pathway and produces leafy galls.Mol. Plant–Microbe Interact. 20, 751–758.

Juhas, M. et al., 2009. Genomic islands: tools of bacterial horizontal genetransfer and evolution. FEMS Microbiol. Rev. 33, 376–393.

Kers, J.A. et al., 2005. A large, mobile pathogenicity island confers plantpathogenicity on Streptomyces species. Mol. Microbiol. 55, 1025–1033.

Kurtz, S. et al., 2004. Versatile and open software for comparing largegenomes. Genome Biol. Available from: <http://genomebiology.com/2004/5/2/R12>.

Lee, C.A. et al., 2007. Identification and characterization of int (integrase),xis (excisionase) and chromosomal attachment sites of the integrativeand conjugative element ICEBs1 of Bacillus subtilis. Mol. Microbiol. 66,1356–1369.

Loria, R. et al., 2006. Evolution of plant pathogenicity in Streptomyces.Annu. Rev. Phytopathol. 44, 469–487.

Martin-Hernandez, A.M. et al., 2000. Effects of targeted replacement ofthe tomatinase gene on the interaction of Septoria lycopersici withtomato plants. Mol. Plant–Microbe Interact. 13, 1301–1311.

McAuliffe, O. et al., 2001. Lantibiotics: structure, biosynthesis and mode ofaction. FEMS Microbiol. Rev. 25, 285–308.

Miyajima, K. et al., 1998. Streptomyces turgidiscabies sp.. Nov. Int. J. Syst.Bacteriol. 48, 495–502.

Nguyen, C.C., Saier Jr., M.H., 1995. Phylogenetic, structural and functionalanalyses of the LacI-GalR family of bacterial transcription factors.FEBS Lett. 377, 98–102.

Ramsay, J.P. et al., 2006. Excision and transfer of the Mesorhizobium lotiR7A symbiosis island requires an integrase IntS, a novelrecombination directionality factor RdfS, and a putative relaxaseRlxS. Mol. Microbiol. 62, 723–734.

Rutherford, K. et al., 2000. Artemis: sequence visualization andannotation. Bioinformatics 16, 944–945.

Scott, J.R., Churchward, G.G., 1995. Conjugative transposition. Annu. Rev.Microbiol. 49, 367–397.

Seipke, R.F., Loria, R., 2008. Streptomyces scabies 87–22 possesses afunctional tomatinase. J. Bacteriol. 190, 7684–7692.

Takeuchi, T. et al., 1996. Phylogenetic analysis of Streptomyces spp.causing potato scab based on 16S rRNA sequences. Int. J. Syst.Bacteriol. 46, 476–479.

te Poele, E.M. et al., 2008a. Actinomycete integrative and conjugativeelements. Antonie Leeuwenhoek 94, 127–143.

te Poele, E.M. et al., 2008b. Actinomycete integrative and conjugativepMEA-like elements of Amycolatopsis and Saccharopolyspora decoded.Plasmid 59, 202–216.

Willey, J.M., van der Donk, W.A., 2007. Lantibiotics: peptides of diversestructure and function. Annu. Rev. Microbiol. 61, 477–501.